The psychobiology of nicotine dependence

It is now widely accepted that nicotine is the principal addictive component of tobacco smoke and that a majority of habitual smokers become dependent upon the nicotine they inhale. As a result, most smokers find it difficult to readily quit the habit in spite of the fact that they know that continuing to smoke is likely to cause significant harm to their health. Most countries in the developed world now seek to diminish the harm caused by tobacco by supporting major programmes directed at promoting smoking cessation. Many of these programmes advocate the use of nicotine replacement therapy (NRT) as an aid to smoking cessation. However, these programmes are often somewhat less successful than anticipated. This observation suggests that the addiction to tobacco may be more complex than a simple dependence upon nicotine. Pre-clinical studies in experimental animals have recently thrown significant light on the reasons why NRT may not provide a complete treatment for the addiction. This article will first consider the evidence that although nicotine has many of the properties of a drug of dependence, its reinforcing properties do not appear to be sufficiently strong enough to explain the powerful addiction to tobacco experienced by many habitual smokers. Some of the factors that may explain this conundrum will then be considered.

THE ADDICTIVE PROPERTIES OF NICOTINE

NEUROBIOLOGICAL MECHANISMS UNDERPINNING NICOTINE DEPENDENCE

THE PUTATIVE ROLE OF MESOLIMBIC DA IN ADDICTIVE BEHAVIOUR

This article had previously highlighted the evidence that mesolimbic DA neurones appear to play important roles in both the reinforcing and rewarding properties of nicotine. However, it has become increasingly clear that the relationship between the effects of nicotine and its withdrawal on behaviour considered relevant to nicotine dependence are more complex than originally thought. There is evidence that the accumbal shell forms part of an extended amygdala and clearly forms part of the limbic system, whereas the accumbal core sends major projections to motor areas of the brain and may be associated more with the control of motor function [35, 36]. Thus, it seems reasonable to suggest that the effects of nicotine on DA overflow in the shell subdivision of the accumbens mediate its rewarding and reinforcing properties, whereas its effects on the DA projections to the accumbal core mediate the locomotor stimulant properties of the drug, which are also enhanced by prior exposure to the drug [18]. This explanation seems consistent with the behavioural data summarised above and with results generated by studies with other drugs, such as amphetamine and methylphenidate [37, 38]. However, it is important to note that the results of some experiments are not entirely consistent with this conclusion. For example, Laviolette and co-workers [39, 40] have reported that micro-injections of nicotine into the ventral tegmental area can elicit an aversive response to nicotine when measured using a conditioned place aversion task. Additionally, there is evidence that the administration of N-methyl-D-aspartate receptor antagonists can block both the development and expression of sensitised DA responses to nicotine measured in the accumbal core, whereas these antagonists do not influence the development or expression of the sensitisation locomotor response to nicotine [41, 42]. Thus, the effects of nicotine on the DA projections to the accumbens do not invariably evoke a rewarding response nor are its effects on DA overflow in the accumbal core invariably tightly associated with increased locomotor activity. Therefore, any hypothesis that seeks to explain the role of mesolimbic DA in nicotine dependence needs to address these controversies.

The DA neurones that project to the nucleus accumbens can fire in different modes, for example as single irregular spikes or as bursts of activity, which are thought to encode different functions and have different consequences [43]. Furthermore, the electron microscope study of Nirenberg et al. [44] suggests that the DA fibres which innervate the nucleus accumbens form two types of varicosity. Some form tight-synaptic contacts with neurones or dendrites within the accumbens. Others form “open synapses” that release DA directly into the interstitial space between the cells sampled by microdialysis probes. Balfour et al. [45] have suggested that single spikes of activity result in a preferential release of DA into the synaptic cleft. This results in activation of post-synaptic DA receptors. Balfour et al. [45] speculated that the DA released from the varicosities that form open synapses occurs preferentially when the neurones switch to the burst firing mode and diffuses into the interstitial space between the cells. By a process of volume transmission, this DA activates receptors on a number of adjacent cells. If this hypothesis is true, it seems reasonable to propose that DA released in response to single spikes may mediate events encoded by the phasic release of DA into the synaptic cleft, whereas DA released by burst firing may serve a paracrine function encoding more tonic events [18, 45]. Thus, for example, it may be possible to explain why both the aversive and reinforcing properties of nicotine may be associated with DA release from the DA neurones, which project to the nucleus accumbens, if it is hypothesised that the effects that the drug exerts on phasic and tonic DA release in different subdivisions of the accumbens subserve different functions and have different behavioural sequelae. This conclusion is largely consistent with the hypothesis proposed by Grace et al. [33], who proposed that synaptic DA mediates events dependent upon the phasic release of DA whereas the DA in the interstitial space between the fibres mediates responses dependent upon the tonic release of DA. However, Grace et al. [33] attribute different functions to burst and spike firing of the neurones. Grace et al. [33] argue that burst firing of the neurones selectively stimulates DA release into the synaptic cleft whereas the tonic release of DA into the interstitial space is mediated by spike firing of the neurones.

One feature that is shared by psychostimulant drugs of abuse, such as amphetamine, cocaine and nicotine, is their ability to evoke a substantial and sustained increase in the concentration of DA in the extracellular space between the cells sampled directly using microdialysis probes. Balfour [11, 18] has argued that it is this shared property which confers addictive potential upon the drugs, and has speculated on the role increased extracellular DA in the accumbal core and shell might play in the neurobiology underlying addiction. There is a substantial body of evidence to support the hypothesis that the DA projections to the nucleus accumbens shell play a central role in the rewarding and reinforcing properties of these drugs when assessed using either the conditioned place preference paradigm of IVSA experiments [11, 18]. Balfour [11] has argued that the changes in the extracellular (putatively paracrine) DA concentration evoked by drugs of dependence in this subdivision of the accumbens are too sustained to be involved in encoding discrete events, such as the presentation of a single discrete reward. It was postulated that the primary role of this DA is to influence the probability that the individual, be it an experimental animal or a smoker, repeats specific behaviours associated with the delivery of a reward. Its primary psychophysiological role is to facilitate the acquisition of behaviours that result in the delivery of rewards, such as food when the individual is hungry. Studies with experimental animals support this hypothesis to the extent that they have shown the delivery of such rewards results in increased DA overflow in the accumbal shell [46]. Both Di Chiara [19, 20] and Balfour [11, 18] have postulated that drugs of abuse exert such large and sustained effects on DA release in the shell that behaviours leading to the delivery of the drug become compulsive and, thereby, dominate the behavioural repertoire of an addicted individual. This implies that an addicted individual comes to derive great pleasure from the behaviour itself through its repeated association with increased DA overflow in the accumbal shell. The hypothesis may explain why, when delivery of the drug is dependent upon the behaviour, the effects of the drug on DA overflow in the accumbal shell are magnified further via conditioning [32, 47]. If the hypothesis is true, it represents an important component of the neurobiology underpinning the development of dependence.

It seems likely that the DA projections to the accumbal core play a role in the mediating elements of the locomotor stimulant properties of psychostimulant drugs [37, 38]. However, the locomotor stimulant properties of nicotine do not appear to be associated directly with changes in the extracellular DA concentration in this region of the brain since pharmacological interventions that block the effects of nicotine on DA overflow in the accumbal core do not invariably result in any changes in the locomotor response to the drug [41, 42, 48]. Balfour [11, 18] has proposed that extracellular DA in this subdivision of the accumbens plays a similar role to that proposed for extracellular DA in the accumbal shell, in that it serves to influence the probability that an individual will exhibit drug-seeking behaviour. However, its primary role appears to be to mediate the effects of sensory and environmental cues on drug-seeking behaviour, although the evidence to date that supports this conclusion is largely circumstantial. Nevertheless, Robinson and Berridge [49, 50] have long argued that the sensitised DA responses evoked by the repetitive administration of addictive drugs influences the incentive salience of cues associated with delivery of the drugs. Furthermore, there is evidence from studies with cocaine that the non-contingent presentation of such a cue leads to both drug-seeking behaviour and a regionally selective increased DA overflow in the accumbal core [51, 52]. Thus, it seems reasonable to hypothesise that the effects of repeated nicotine exposure on DA overflow in the accumbal shell and core exert complementary effects on nicotine-seeking behaviour which, together, serve to enhance greatly the addictive properties of the drug [11].

THE ROLE OF CONDITIONED STIMULI IN NICOTINE DEPENDENCE

Although nicotine appears to have the behavioural and neurobiological properties of a drug of dependence, its reinforcing properties in experimental rats are relatively weak when compared with those of many other drugs of abuse and do not appear to provide an adequate explanation for the powerful addictive properties of tobacco smoke [11, 53]. There are a number of reasons why nicotine may not be a powerful reinforcer in this species. Le Foll et al. [54] have reported data which suggest that nicotine is more reinforcing in higher animals, such as monkeys, with complex cognitive skills, which are lacking in rodents. There is also evidence that tobacco smoke contains other, non-nicotinic compounds, which interact pharmacologically to enhance the neurobiological and behavioural responses to nicotine. For example, tobacco smoke contains compounds that inhibit the enzyme, monoamine oxidase, which metabolises DA and could, therefore, be expected to enhance the responses to nicotine mediated through DA [55]. There is, indeed, experimental evidence using animal models that inhibition of monoamine oxidase enhances the reinforcing properties of nicotine when measured using the IVSA paradigm [56].

Another important factor may be the fact that prolonged or sustained exposure to nicotine causes desensitisation of many of the nAChRs, which mediate its effects in the brain [11, 45]. As a result, while the administration of nicotine to animals with low circulating levels of nicotine results in stimulation of nAChRs, its administration to animals with blood nicotine concentrations similar to those found in the plasma of habitual smokers for much of the day fails to stimulate many of DA neurones in the ventral tegmental area or result in the release of DA in the nucleus accumbens [57, 58]. Balfour and co-workers [11, 18, 25, 45] have argued that the relatively limited reinforcing properties of nicotine are related to the fact that sustained exposure to the drug causes desensitisation of nAChRs, thereby restricting the circumstances in which nicotine stimulates DA overflow in the nucleus accumbens. In this sense, the effects of nicotine contrast with those of other psychostimulant drugs of abuse which exert their effects on DA overflow through different molecular mechanisms and which persist for the full period of drug administration. Thus, any hypothesis that seeks to explain the role of nicotine in the addiction to tobacco must take into account the fact that, for many smokers, the inhalation of nicotine in tobacco may not necessarily result in the stimulation of the DA projections to the nucleus accumbens, which appears to be pivotal to the development of nicotine dependence.

In experimental models, the self-administration of psychostimulant drugs of abuse is facilitated significantly if delivery of the drug is paired with a sensory stimulus, such as a light and/or a tone. A series of studies by Caggiula and co-workers [59–61] have shown that the incorporation of this type of stimulus also greatly facilitates the acquisition and maintenance of nicotine self-administration. Indeed, in studies with experimental rats it often proves difficult to demonstrate significant levels of nicotine self-administration without using these stimuli. Importantly, once the association between a stimulus and nicotine delivery has been established, responding can be maintained, albeit at a lower level, by these stimuli in the absence of nicotine [59]. Furthermore, in rats in which responding for nicotine has been extinguished by replacing the drug reinforcer with an injection of saline, the presentation of the conditioned stimulus evokes nicotine-seeking behaviour [53, 59, 61, 62]. This paradigm is commonly used to model the way in which cues associated delivery of an addictive drug precipitate relapse in an abstinent individual. Thus, these experiments suggest that conditioned stimuli can play an important role in nicotine self-administration in experimental animals.

Studies performed in human smokers imply that they probably play an equally important role in the addiction to tobacco. For example, Rose et al. [63] have reported that, in established smokers, sensory-motor cues present in tobacco smoke play a role in regulating the tobacco smoking habit and ameliorating the craving to smoke experienced by abstinent smokers. A more recent study showed that many habitual smokers found nicotine-free cigarettes almost as satisfying as regular cigarettes and that inhaling the smoke from these denicotinised cigarettes relieved the craving to smoke [64]. The sensory stimuli present in tobacco smoke seem to be particularly important in highly addicted smokers [65]. It seems reasonable to propose, therefore, that the powerful addiction to tobacco reflects an interaction between a relatively weak reinforcing drug, nicotine, and the sensory cues present in the tobacco-smoke vehicle in which it is delivered [18, 66].

The nature of the interaction between the sensory motor cues associated with the delivery of nicotine and “satisfaction” derived by the smoker when they inhale tobacco smoke may be more complex than was initially thought. Sensory motor cues may serve as “classical” conditioned stimuli in which their efficacy as conditioned reinforcers depends entirely upon their close association with behaviours that deliver the drug (e.g. smoking in humans, lever-pressing in rats trained to respond for nicotine). In the absence of such conditioning, these stimuli do not have any intrinsic reinforcing properties. Recent experiments have provided evidence that association with the delivery of nicotine can confer reinforcing properties on stimuli of this type [53, 62]. However, nicotine also appears to have the ability to significantly enhance the reinforcing properties of an otherwise weakly reinforcing stimulus. This has been demonstrated using a paradigm in which rats were trained to switch off the house light in an operant chamber (rats prefer low light environments). In this paradigm, control animals can be trained press a lever to extinguish the light, although this reward does not have powerful reinforcing properties [61]. In rats treated with nicotine, responding for this stimulus is greatly enhanced. However, Donny et al. [61] showed that the effect of nicotine in this paradigm does not depend upon a close association between a self-administered i.v. injection of the drug and the presentation of the stimulus. Indeed, in a yoked design, animals who received the nicotine injections non-contingently (i.e. they did not control when they received the injections) pressed just as often as the rats in which presentation of the sensory stimulus was tightly associated with the contingent self-administration of nicotine. These data imply that the rats were responding primarily for the sensory stimulus (extinguishing the house light) and that nicotine greatly enhanced the reinforcing properties of the stimulus irrespective on the dependency with which it was delivered.

In the study by Donny et al. [61], small i.v. injections of nicotine were delivered throughout the 1-h trials. A recent pilot study in the current authors’ laboratory has demonstrated that a single subcutaneous injection of the drug, delivered 10 min before the trial, exerts a very similar effect on responding reinforced by extinguishing the house light (fig. 1⇓). These results clearly support those of Donny et al. [61], but also suggest that a single injection of nicotine exerts effects in the brain that persist long enough to facilitate responding for the stimulus for at 1 h. Nicotine exerts a number of effects in the brain which could mediate this effect. Interestingly, however, the non-contingent administration of amphetamine, at a dose that elicits a sustained increase in DA overflow in the nucleus accumbens, facilitates responding for a conditioned stimulus previously paired with another reinforcer [67, 68]. Therefore, it seems reasonable to speculate that the sustained effects that nicotine exerts on DA overflow in the accumbens may also mediate its effects on responding for a weak reinforcer; however, other possible explanations should not be excluded.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

The effects of non-contingent nicotine on responding for a weak reinforcer. Responses in an operant chamber on a) the active lever, which turned out the house light for 20 s, and b) the inactive lever, which had no programmed consequences. The rats were given a subcutaneous injection of saline (□) or nicotine (▓; 0.4 mg·kg⁻¹) 10 min before being placed in the operant chamber. Responses were measured over the course of the next 60 min. Response was measure on a fixed ratio (FR) schedule of reinforcement with a 20-s time-out between the availability of the reward. Data are reported for responding on an FR1, FR2 and FR5 schedule in which the rats were required to press once, twice or five times for the reinforcer. Pressing on the active lever was significantly higher than the inactive lever (F(1,20) = 28.1, p<0.001). In the rats treated with nicotine but not saline, the increase in response seen on the active lever as the FR increased was significant (F(2,20) = 9.4; p<0.001). This was significantly different from rats treated with saline. Data are expressed as mean±sem of six observations. *: p<0.05; **: p<0.01. (N. Montgomery and D.J.K. Balfour; unpublished data).

THE IMPLICATIONS OF LABORATORY STUDIES FOR OUR UNDERSTANDING OF PSYCHOPHARMACOLOGY AND TREATMENT OF TOBACCO DEPENDENCE

The laboratory studies summarised above have shown that nicotine has behavioural and neurobiological properties characteristic of a drug of dependence and provide clear support for the primary role of nicotine in the addiction to tobacco. Thus, they also provide support for the use of NRTs as an aid to smoking cessation. NRT clearly elicits a significant and clinically valuable increase in the number of smokers who are able to successfully quit their habit [69, 70]. However, the pharmacology underlying the effectiveness of this treatment may be more complex than a simple substitution therapy. A majority of NRT preparations deliver nicotine to the brain more slowly than cigarette smoke and it seems plausible that the main therapeutic effect of NRTs reflects their ability to maintain plasma nicotine levels at a sufficiently high concentration to limit the development of nicotine withdrawal [11]. More recently, varenicline, a drug that acts as a partial agonist at the α₂β₄ nAChR, has been introduced as an aid to smoking cessation [71]. In many ways, this compound mimics the pharmacological properties which are seen as being important to addiction, i.e. the ability to both stimulate and block the α₂β₄ nAChR. At the experimental level, the administration of this compound to drug-free experimental rats increases DA overflow in the nucleus accumbens but blocks the effects of a subsequent challenge with nicotine [72]. Furthermore, it inhibits nicotine self-administration and attenuates the reinstatement of nicotine-seeking behaviour evoked by a non-contingent injection of nicotine [72]. Thus, it might be expected to limit the effects of nicotine withdrawal while at the same time preventing a challenge dose of nicotine exerting effects in the brain which mediate reinforcement. The hypothesis would seem to hold true to the extent that the drug is clearly an effective aid to smoking cessation with an efficacy at least as great, and possibly greater than, NRT [73].

Another non-nicotinic approach that has proved valuable is bupropion. The efficacy of bupropion as an aid to smoking cessation was discovered serendipitously and the psychobiology underlying its efficacy in this indication remains the subject of debate. This compound, or its active metabolites, potentiate the effects of norepinephrine and DA in the brain by blocking the transporters which re-accumulate these neurotransmitters into the nerve terminals, and it is assumed that this effect is important to its mechanism of action as an aid to smoking cessation [74]. If this is true, it seems reasonable to hypothesise that the compound works by alleviating some of the consequences of nicotine/tobacco withdrawal through its effects on these pathways, perhaps most notably the anhedonia experienced by many abstinent smokers. Recent studies with experimental animals provided some support for this hypothesis, to the extent that the compound attenuates the anhedonia measured in nicotine abstinent rats [75]. Bupropion also attenuates the somatic consequences of nicotine withdrawal [76]. Interestingly, in animals allowed to self-administer nicotine, bupropion did not reduce responses for the drug although it did diminish the increase in brain reward function exhibited by these animals, suggesting a dissociation between these two responses to the drug [76].

Although the use of NRT results in a significant and clinically valuable increase in the number of smokers who are able to quit successfully, this approach is often not as successful as therapists and smokers would like. The results of the studies with experimental animals have, perhaps, provided an explanation for this conundrum. The experimental data suggest that nicotine may not be an intrinsically powerful reinforcer when compared with many of the other drugs of abuse used by humans. This article has summarised some of the reasons why this may be true and considered some of the reasons why a drug that is a relatively weak reinforcer may, nevertheless, be of fundamental importance to the powerful addiction to tobacco smoke experienced by many smokers. It is possible that the addiction reflects an interaction between the reinforcing properties of the drug and the complex cognitive skills of humans relative to the experimental animals in which most of the behavioural studies have been performed [54]. However, it also now seems likely that the sensory stimuli within the tobacco smoke vehicle are also of primary importance and the addiction to tobacco reflects a psychobiological interaction between these stimuli and the effects of nicotine in the brain [11]. Thus, for many smokers the dependence upon tobacco may usefully be equated to a second order schedule of reinforcement in which the reinforcing properties of nicotine per se are experienced relatively infrequently. For the periods when the nAChRs are desensitised, which may be significant parts of the day, the habit is reinforced by cues and conditioned stimuli present in the smoke [11]. The hypothesis is summarised in figure 2⇓. This possibility impacts significantly on the approaches that need to be adopted to treat tobacco dependence more successfully than it is at present.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Tobacco smoking as a second order schedule of reinforcement. Schematic diagram showing the postulated differential dopamine (DA) responses to nicotine in tobacco smoke when delivered following a period of abstinence and following a period of continued smoking. This causes an accumulation of nicotine in the blood and brain resulting in desensitisation of neuronal nicotinic receptors. The figure demonstrates the proposed role of nicotine, conditioned sensory stimuli within tobacco smoke and avoidance of withdrawal as reinforcers of the tobacco smoking habit in nicotine-abstinent and nicotine-replete individuals.

It seems reasonable to propose that nicotine-based therapies preferentially target the nicotinic component of the addiction. This is probably also true of other therapies, such as varenicline, which depend upon their effect on nicotinic systems within the brain. Drugs which antagonise these receptors have been shown to attenuate the rewarding properties of nicotine, measured using the CPP paradigm [77, 78], and the reinforcing properties of the drug assessed using nicotine IVSA [79, 80]. Interestingly, however, the administration of CB1 receptor antagonists attenuates the persistent nicotine-seeking behaviour reinforced by the presentation of conditioned stimuli previously associated with the delivery of nicotine [62] and the reinstatement of nicotine-seeking behaviour evoked by the non-contingent presentation of a priming dose of nicotine or a conditioned stimulus [80]. These observations support the conclusion that endocannabinoid systems within the brain play a role in the neurobiology underpinning nicotine dependence and suggest that drugs targeting the CB1 receptors which mediate their effects may be useful in the treatment of tobacco dependence. Importantly, these compounds seem to target a system putatively implicated in the mechanisms that mediate the role of conditioned stimuli in the addiction [81].

In summary, therefore, the evidence that nicotine is the primary addictive component of tobacco remains robust. However, studies in recent years suggest that the tobacco smoke vehicle in which it is delivered also plays a pivotal role in the addiction to tobacco. Tobacco smoke contains other pharmacologically active agents that can enhance or facilitate the dependence upon nicotine. Additionally, it would seem that the dependence upon nicotine is facilitated by the close association with sensory cues within the smoke, which both greatly enhance nicotine-seeking behaviour and can, themselves, maintain the smoking habit. Future improvements in the treatment for tobacco dependence and relapse will need to address this component of the addiction.

Statement of interest

D.J.K. Balfour has received an honorarium of US$2,500 for speaking at a symposium organised by Pfizer, a fee of US$2,500 for serving on a grant panel funded by Pfizer and a consultancy fee from Sanofi-Aventis for attending a research meeting. Research sponsorship was also received from Rothmans UK for two PhD students and a Postdoctoral researcher for approximately 7 years, this support ceased around 1990. D.J.K. Balfour's current work on nicotine and tobacco dependence is supported by grants from the Wellcome Trust and Cancer Research UK.